A coupled stochastic rainfall-evapotranspiration model 1 for hydrological impact analysis 2 Minh Tu Pham *1 , Hilde Vernieuwe 2 , Bernard De Baets 2 , and Niko E. C. Verhoest 1 3 1 Laboratory of Hydrology and Water Management, Ghent University, Coupure 4 links 653, 9000 Ghent, Belgium 5 2 KERMIT, Department of Mathematical Modelling, Statistics and Bioinformatics, 6 Ghent University, Coupure links 653, 9000 Ghent, Belgium 7 Abstract 9 A hydrological impact analysis concerns the study of the consequences of certain scenarios 10 on one or more variables or fluxes in the hydrological cycle. In such exercise, discharge is often 11 considered, as especially extreme high discharges often cause damage due to the coinciding 12 floods. Investigating extreme discharges generally requires long time series of precipitation and 13 evapotranspiration that are used to force a rainfall-runoff model. However, such kind of data 14 may not be available and one should resort to stochastically-generated time series, even though 15 the impact of using such data on the overall discharge, and especially on the extreme discharge 16 events is not well studied. In this paper, stochastically-generated rainfall and coinciding 17 evapotranspiration time series are used to force a simple conceptual hydrological model. The 18 results obtained are comparable to the modelled discharge using observed forcing data. Yet, 19 uncertainties in the modelled discharge increase with an increasing number of stochastically- 20 generated time series used. Notwithstanding this finding, it can be concluded that using 21 a coupled stochastic rainfall-evapotranspiration model has a large potential for hydrological 22 impact analysis. 23 1 Introduction 24 Precipitation is the most important variable in the terrestrial hydrological cycle that determines 25 soil moisture and discharge from a watershed. As such, it also impacts water management where 26 generally the occurrences of extreme events, e.g. storms or droughts, which have very low frequen- 27 cies, are of concern. Very long time series of precipitation are hence needed. Because this kind of 28 data is not always available, one may consider using a stochastically-generated rainfall time series 29 (Boughton and Droop, 2003). Stochastic rainfall models can be used to produce very long time 30 series or to compensate for missing data from finite historical records (Wilks and Wilby, 1999). 31 Several types of rainfall models have been proposed in literature. Onof et al. (2000) grouped all 32 continuous rainfall models into four types: (1) meteorological models; (2) stochastic multi-scale 33 models; (3) statistical models and (4) stochastic process models. Meteorological models are capa- 34 ble to describe the physical processes of all weather variables, including rainfall, by making use of 35 very large and complex sets of equations. Numerical Weather Prediction and General Circulation 36 Models are two common examples of this type of models. Stochastic multi-scale models describe 37 the spatial evolution of the rainfall process regardless of scale factors. In general, these models 38 involve an assumption of temporal invariance of rainfall over a range of scales (Bernardara et al., 39 2007). Statistical models, which can be used for simulating the precipitation trends, usually treat 40 * MinhTu.Pham@UGent.be 1 8 Hydrol. Earth Syst. Sci. Discuss., doi:10.5194/hess-2017-161, 2017 Manuscript under review for journal Hydrol. Earth Syst. Sci. Discussion started: 23 May 2017 c  Author(s) 2017. CC-BY 3.0 License.